This study characterised the growth, nutritional status and management of CD and T1DM in children with dual diagnoses before and after the diagnosis of CD by comparison with a matched cohort of CD and T1DM controls. Such aspects make this study unique and overcome the limitations of previous research [7, 11, 12, 14, 16] (Additional file 1: Table S1). In summary, the findings of this study suggest that the dual diagnoses of CD and T1DM do not affect the nutritional status, growth and disease management of either condition. This is in accordance with previous studies which reported no difference in the growth patterns of children with T1DM + CD at CD diagnosis compared with their peers with T1DM (14–16) but contrary to others, who found deterioration of linear growth [7, 17], weight [7, 17, 22, 23] and BMI z-scores  both at and post CD diagnosis (Additional file 1: Table S1). Inconsistency between studies may be explained by differences in screening practices between health centres, delayed diagnosis of CD and variations in sample size or sample selection bias by comparing groups of participants with different characteristics (Additional file 1: Table S1). Indeed, studies that reported similar results to the current study were located in centres which had annual screening programs for CD in people with T1DM. In studies where CD screening was not routine practice, height and weight z-scores deficits were observed in children with T1DM + CD compared with T1DM controls [16, 17].
Weight gain is a recognised feature in individuals with T1DM on intensive insulin therapy . T1DM is often characterised by weight loss prior to diagnosis of diabetes and weight gain following initiation of insulin therapy. This is a confounder in previous studies that did not account for fluctuations in growth at the start of insulin therapy [11, 13, 17]. In the current study, all children with T1DM and T1DM + CD were specified to have been diagnosed with T1DM diagnosis more than two years prior to CD diagnosis to eliminate the remission phase after T1DM diagnosis. Moreover as the T1DM control group was also matched for age at T1DM diagnosis and duration of T1DM at CD onset (baseline) any influence would be reflected evenly in both groups.
Our results are consistent with many studies reporting no significant difference in HbA1c levels between individuals with T1DM and T1DM + CD before or after the diagnosis of CD [12, 13]. However, other studies have reported lower HbA1c levels [14, 16] or raised HbA1c levels , along with a reduction in insulin requirements and an increase in the number of hypoglycaemic episodes in the period prior to CD diagnosis [11, 12]. This was attributed by the authors to a possible destruction of the intestinal mucosa impinging nutrient absorption . In the current study, HbA1c levels were raised significantly one and two years post CDx in the T1DM + CD children, however a similar increase was observed in those with T1DM which suggests that this is a feature of T1DM disease progression rather than a result of CD diagnosis. This may be due to adolescents presenting poorer adherence to insulin therapy  and a large proportion of T1DM + CD are diagnosed during this age.The absence of significant fluctuations in HbA1c levels in the children with T1DM + CD in our study may be a result of the regular screening practices, predisposing to the earlier diagnosis of CD, but also due to the high level of dietetic supervision that these children receive in our centre.
Strict compliance to a GFD is associated with mucosal histology restoration and improved nutritional status in people with CD but adherence is difficult to assess accurately and has been frequently over-looked in studies [7, 11, 14, 15]. In this study a significantly lower proportion of T1DM + CD children presented with normal tTg levels (<7.0 U/ml) one year post CD diagnosis than those with CD; by two years the majority of children in both groups had tTg levels within the normal values. This may reflect that people with T1DM + CD need more time to adapt to the challenges and dietary management of both conditions particularly as the large majority were diagnosed during adolescence, when adherence to GFD is reported between 25-60% [18, 19]. On the other hand, active on-going symptoms in those with CD may encourage adherence to a GFD making the dietary changes more acceptable. Moreover, individuals with T1DM + CD identified by serological testing with “silent” CD have an absence of overt clinical symptoms, therefore the GFD may be perceived as of secondary priority to the management of diabetes and its associated complications.
One year prior to CD diagnosis, the mean change in BMI z-scores in the children with T1DM tended to increase but in those with T1DM + CD, the average change in BMI z-scores were significantly lower and negative. These differences in nutritional status change may indicate the onset of CD before diagnosis, which might have affected nutrient absorption and subsequently nutritional status. This pattern was reversed within two years after the introduction of GFD, supporting our speculations. Moreover it is possible that further deterioration might have been prevented by routine screening and early detection of CD in these people, limiting the time CD would remain undiagnosed. In children with CD it is well established that the longer active CD remains undiagnosed and the older the age before this is diagnosed are both associated with complications like gastrointestinal malignancy  and osteoporosis [27, 28]. Yet in T1DM the long-term effects of sub-clinical CD are not well investigated and remain grossly unknown. It has been speculated that undetected and untreated CD may compound the effects of T1DM and increase the risk of the developing of CD associated complications, delaying the benefits to nutritional status and psychological wellbeing associated with GFD introduction [1, 29].
The retrospective design of this study is an inherent limitation as often leads to incomplete datasets and sample selection bias Moreover, data on insulin requirements and number of hypoglycaemic episodes, which would reflect a better assessment of T1DM management, were not available.
High quality longitudinal studies are required to understand the long-term outcomes of T1DM + CD. Due to lack of evidence the National Institute of Clinical Excellence (NICE) in the UK revised its guidelines reducing routine screening to a single screen at T1DM diagnosis . Evidence from this study suggests that regular screening may have prevented the further decline in nutritional status and tackled deterioration of growth velocity in people with silent T1DM + CD.